Ad signal

In Saudi Arabia, iOS adoption rates are among the highest in the region. A meaningful share of real conversions were disappearing before they ever reached Meta. The platform was making delivery decisions with partial data, finding users who were cheap to reach, not users likely to buy.

The brand raised membership prices in 2025. Higher prices create friction. Checkouts grew anyway. That is Snapchat finding better buyers, not just more buyers.

Spend was down 50%. ROAS nearly tripled. Once Meta had a complete picture of who was actually buying, the budget that remained went further than the full budget had before.

Before Journify, purchases were down 37% year on year and cost per purchase was up 70%. The audience hadn't changed. The signal had. Meta was optimizing with an incomplete picture of who was actually moving toward purchase.

Snapchat was receiving incomplete checkout data and optimizing around the wrong signals. Match rates were below what the platform needed to learn. Once cleaner signals arrived, the algorithm figured out who was actually buying.

Snap's algorithm optimizes on what it sees. When Lumi fixed what it was sending, everything else followed.

Baytonia's TikTok campaigns were running on incomplete purchase data. The algorithm was learning from a partial picture of who was actually buying. Once the full signal reached the platform, targeting sharpened and cost came down.

Jarir's campaigns were generating real purchases across Saudi Arabia and Kuwait that Meta couldn't fully see. The platform was allocating budget based on a partial picture of who was actually buying.

Snapchat was learning from half the picture. Server-side events through Journify were reaching the platform at twice the volume of pixel-only events. Once the full signal arrived, the algorithm stopped optimizing on noise and started finding real property buyers.

Dr. Nutrition's Snapchat campaigns were generating real purchase activity the platform couldn't fully see. Cost per purchase was elevated and click quality was limited. The campaign wasn't the problem. The data feeding it was.

Qasr Al Awani's Snapchat campaigns were running on an Event Quality Score of 4.3 out of 10. The platform's AI was learning from incomplete data. Fixing the signal changed what the algorithm could see and who it could find.

Al Hokail's Snapchat campaigns were generating real sign-up activity the platform couldn't fully see. Incomplete signals meant the algorithm was targeting on guesswork. Fixing the data layer changed what Snapchat could optimize toward.